|Home | About | Journals | Submit | Contact Us | Français|
Relatively little is known about factors that contribute to the development of meningioma and vestibular schwannoma, two intracranial nervous system tumors. We evaluated the risk of these tumors in relation to family history of malignant or benign tumors. Incident cases of meningioma (n = 197) or schwannoma (n = 96) were identified at three U.S. referral hospitals between June 1994 and August 1998. Controls (n = 799) admitted to the same hospitals for nonmalignant conditions were matched to cases on age, sex, race/ethnicity, hospital, and proximity of residence to hospital. We found that risk of meningioma was increased among persons reporting a family history of a benign brain tumor (odds ratio [OR], 4.5; 95% confidence interval [CI], 1.0–21.0; n = 5) or melanoma (OR, 4.2; 95% CI, 1.2–15.0; n = 5). A family history of breast cancer was associated with an elevated meningioma risk among participants aged 18 to 49 years (OR, 3.9; 95% CI, 1.4 –11.0; n = 8) but a reduced risk among older respondents (OR, 0.2; 95% CI, 0.1–0.7; n = 3). Family history of cancer did not differ between schwannoma cases and controls, although the statistical power to detect associations was limited. Some relative risk estimates were based on a small number of observations and may have arisen by chance. Inheritance of predisposing genes, shared environmental factors, or both within families with a history of benign brain tumors, melanoma, or possibly breast cancer may be related to altered meningioma risk.
Meningioma and vestibular schwannoma are intracranial nervous system tumors that are predominantly benign in behavior and of uncertain etiology. Only a small fraction of meningiomas (<10%) are malignant (McCarthy et al., 1998), but even among the large majority of benign meningiomas, a proportion may recur (7%–30%) (McCarthy et al., 1998; Mirimanoff et al., 1985), necessitating repeated brain surgery. Vestibular schwannomas, also known as acoustic neuromas, are distinguished from other nervous system tumors by their intracranial origin in the Schwann cells surrounding the eighth cranial nerve. Their benign nature belies the considerable morbidity that can accompany these neoplasms, most notably, permanent hearing loss (Harner and Laws, 1981). Investigations of risk factors for these tumors are rare, in part because historically they have not been included in most cancer registries. One of the few established environmental risk factors for both meningioma and schwannoma is exposure to ionizing radiation, as each tumor occurs in excess among children irradiated for tinea capitis (Ron et al., 1988) and among the survivors of the atomic bomb explosions at Hiroshima and Nagasaki (Preston et al., 2002). The occurrence of these tumors in rare heritable cancer syndromes has provided some initial understanding of the contribution of genetic factors to tumor etiology. Increased risks of meningioma have been noted in neurofibromatosis type 2 (NF2)2 (Wertelecki et al., 1988), hereditary non-polyposis colon cancer (Aarnio et al., 1999), and Werner syndrome (Goto et al., 1996), while schwannoma is one of the defining characteristics of NF2 (Young et al., 1970). In addition to these rare cancer syndromes, the increased risk of meningioma and perhaps schwannoma among Jewish individuals also raises the possibility of an underlying genetic predisposition (Inskip et al., 2003; Preston-Martin 1989).
Studies of family history of cancer can contribute to our understanding of the role of genetic factors in meningioma or schwannoma etiology, and these histories may be particularly useful when previous information is sparse. In four prior case-control investigations, meningioma risk was not elevated among individuals with a family history of meningioma, of specific types of cancer, or of any cancer (Choi et al., 1970; Preston-Martin et al., 1980, 1989; Ryan et al., 1992), but fewer than 100 cases were included in all but one study (Preston-Martin et al., 1980). Larger cancer registry–based record-linkage studies reveal a 2.0- to 2.5-fold increased risk of meningioma or schwannoma among persons with a family history of the same tumor (Hemminki et al., 2001a; Malmer et al., 2003) or any benign or malignant brain tumor (Hemminki et al., 2001a).
We analyzed data to determine whether the family history of benign brain tumors or cancers differed between meningioma or schwannoma cases and controls in a hospital-based case-control study.
The study design has been described in detail previously (Inskip et al., 1999). All individuals diagnosed at three referral hospitals for the treatment of intracranial nervous system tumors (Brigham and Women’s Hospital, Boston, Mass.; St. Joseph’s Hospital, Phoenix, Ariz.; and Western Pennsylvania Hospital, Pittsburgh, Penn.) between June 1994 and August 1998 with incident intracranial meningioma (disease codes 9530–9538 in International Classification of Diseases for Oncology, 2nd edition [ICDO-2; Percy et al., 1990]) or vestibular schwannoma (ICDO-2 9560, 9570) were eligible for the study. Additionally, participants were required to be 18 years of age or older, reside within 50 miles of the admitting hospital (or within the state of Arizona for St. Joseph’s Hospital in Phoenix), speak either English or Spanish, and receive care at a study hospital within eight weeks of the qualifying diagnosis. Individuals with a prior history of a brain or nervous system tumor were excluded. More than 90% of the patients with meningioma and 82% of the patients with schwannoma (excluding proxy interviews) were interviewed within three weeks of diagnosis. All meningioma and 96% of schwannoma diagnoses were confirmed microscopically, with the remaining 4% confirmed by MRI or CT. Overall, 98% of eligible individuals with meningioma and schwannoma agreed to participate.
Controls were selected from individuals admitted to the study hospitals during the same time period who were diagnosed with nonmalignant conditions and were matched to meningioma and schwannoma cases (and to 489 glioma cases described in a separate report [Hill et al., 2003]) on age (10-year age group), race/ethnicity (non-Hispanic white, Hispanic white, African-American, Asian, or other), sex, and distance from residence to hospital (0–5 miles, >5–15 miles, >15–30 miles, >30 miles) so that the distribution of these variables would be approximately equal between all cases and controls. Of eligible controls who were approached for an interview, 799 (86%) agreed to participate, and 92% were interviewed within three weeks of hospital admission. The most common discharge diagnosis categories for controls were trauma, injury, or poisoning (25%) or diseases of the circulatory (22%), musculoskeletal (22%), digestive (11%), or nervous (7%) systems. The study protocol was approved by Institutional Review Boards at the three study hospitals and the U.S. National Cancer Institute. Written informed consent was obtained from each participant or from his or her proxy before the interview.
A trained research nurse administered a structured computer-assisted interview to participants while in the hospital. Interviews were conducted with a proxy (usually a spouse) for 8% of meningioma cases, 1% of schwannoma cases, and 3% of controls, usually because of severe impairment of the respondent, and an additional 6% of meningioma, 1% of schwannoma, and 1% of controls interviews were conducted with both the participant and proxy. The interviewer sought information on demographics, known or suspected risk factors for nervous system tumors, and the occurrence of primary cancer or benign intracranial tumors in first-degree relatives (biological parents, full siblings, and children). A show card list, including a category for “Other—specify” was used to ascertain the type of brain tumor in a family member for later classification as benign or malignant: All reported types of glioma and neuroepitheliomatous tumors corresponding to ICDO-2 morphology codes 9380–9523 were treated as malignant. All reported cases of meningioma, neurofibroma, neuroma, or schwannoma were considered benign, and those that could not be specified were categorized as “Unknown if benign or malignant.” For each blood relative, the interviewer asked about birth date or current age, vital status, and, if applicable, the tumor site and type and the age at diagnosis. The study participant’s personal history of cancer and previous benign tumors (other than of the brain) and personal or family history of neurofibromatosis also was recorded.
A family history of cancer is dependent upon the family size and age of relatives of the respondent, and it can be a biased estimator of familial cancer aggregation, particularly if those factors differ between cases and controls (Khoury and Flanders, 1995). Analytic methods that compare observed cancers among family members to those expected on the basis of the age, sex, and birth cohort of individual relatives, such as calculation of standardized incidence ratios (SIRs), can reduce such bias. Thus, we computed SIRs, using methods that have been described previously (Hill et al., 2003), and compared them to odds ratios (ORs).
We calculated ORs and 95% confidence intervals (CIs) as estimates of the relative risk of brain tumors among persons who reported the occurrence of a specific cancer in a relative, compared with those that did not, using conditional logistic regression. We adjusted for the matching variables (race/ethnicity [collapsed to white/nonwhite to avoid empty cells], sex, distance from residence to hospital [0 –15 miles, >15 miles], and age at interview as a continuous variable) and excluded respondents who could not provide information about first-degree relatives (4 meningioma cases, 3 schwannoma cases, 31 controls). Thus, our results are based on 193 meningioma cases, 93 schwannoma cases, and 768 controls. Potential confounders considered in the analysis were the education and income of the eligible participant, number of first-degree relatives, and the age of parents and two oldest siblings, as these are the relatives most likely to contribute to a family history of cancer. Separate analyses were undertaken according to the age of the respondent (<50 years, ≥50 years), the relative’s age at cancer diagnosis (<50 years, ≥50 years), and type of respondent (self, proxy/assisted). Adjustment for education, income, or number or age of first-degree relatives did not appreciably alter the OR estimates, and thus these variables were not included in the final logistic regression models.
Meningioma and schwannoma cases had a significantly higher income than controls, and schwannoma cases also had a significantly higher level of education (Table 1) (Inskip et al., 2003). Current ages of parents of meningioma cases and controls were similar; however, fathers and mothers of schwannoma cases were significantly older than control parents. As the OR estimate did not differ from the ratio of the SIR estimates in cases and controls, only the adjusted OR risk estimates are presented.
Risk of meningioma was not increased among persons who reported a first-degree relative with a malignant brain tumor, although risk was elevated more than 4-fold in those who reported a family member with a benign brain tumor or melanoma (Table 2), compared with those who did not report such a family history. Persons reporting a family history of liver or gallbladder cancer, prostate cancer, or Hodgkin disease had somewhat increased risks, but CIs were wide and included 1.0. Only five cancer sites (stomach, colon, lung, breast, and prostate) were reported sufficiently often in relatives of schwannoma cases and controls to estimate relative risks, and the proportion with a family history of each cancer was similar in the two groups (Table 2). Further analyses of schwannoma risk according to age of participant or age at cancer diagnosis were precluded by the small numbers of observations.
Among participants aged 18 to 49 years, risk of meningioma was increased 4- to 5-fold in those who reported a family history of breast cancer or melanoma, compared with those who did not report the respective family history (Table 3). In this age group, cases also were more likely than controls to report any malignancy among first-degree relatives, but the difference was not statistically significant. Among older respondents (age ≥50 years), meningioma risk was decreased in persons who reported breast cancer in a first-degree relative but was not altered among those reporting a family history of cancers at other sites. Strong differences in risk were not evident in analyses that examined meningioma risk according to the age at cancer diagnosis of the family member (<50, ≥50) or by respondent type (self, proxy/assisted).
A personal history of cancer was not related to risk of meningioma or schwannoma, but too few participants reported cancer at specific sites for meaningful analysis (data not shown). Risk of meningioma was somewhat increased among persons with a personal history of an additional benign tumor (other than of the brain) (OR 2.3; 95% CI, 1.1–4.6; n = 16), but further investigation was precluded by the limited number of particular tumor types (primarily lipomas, leiomyomas, benign breast tumors, and ovarian fibroids) (data not shown). One meningioma case, two schwannoma cases, and no controls reported a personal history of neurofibromatosis, and one additional schwannoma case reported a family history of neurofibromatosis. None of these individuals reported a family history of brain cancer or benign brain tumors. Respondents were not queried about other rare genetic syndromes that confer an increased risk of meningioma or schwannoma.
In this investigation, meningioma risk was elevated at least 3- to 4-fold among participants with a family history of benign brain tumors, melanoma, and, among those age 18 to 49 years, breast cancer. We did not identify any family history of cancer that was related to schwannoma risk, but had relatively little power to do so because of the small number of cases.
Only a few previous studies have investigated meningioma or schwannoma risk in relation to family history of cancer. Risk of meningioma was not elevated among those with a family history of meningioma, or of all cancers combined, in several case-control studies (Choi et al., 1970; Preston-Martin et al., 1980, 1989; Ryan et al., 1992). Risks of meningioma and schwannoma also have been evaluated by linking genealogical information to a Swedish population-based cancer registry, using methods that accounted for family size and age of relatives. In one such study, risk of meningioma was increased 2.2-fold among those who had a first-degree relative with meningioma (Malmer et al., 2003), and in a second study, risk of adult meningioma was increased 2.5-fold and 1.7-fold, respectively, among those who had a parent with meningioma or an endocrine cancer (Hemminki et al., 2001a). Also, risk of adult schwannoma was elevated 1.5-fold among those who had a parent with any brain tumor (benign or malignant) and 3-fold among those who had a parent with thyroid cancer (Hemminki et al., 2001a).
Our meningioma results are generally consistent with previous reports concerning rare cancer syndromes, multiple cancer families, and second primary cancer studies. Meningioma co-occurs more frequently than expected with other benign nervous system tumors, primarily schwannoma (Wertelecki et al., 1988), in NF2 families. Meningioma and schwannoma have been noted in multiple-case melanoma families that also include brain cancers (Bahuau et al., 1998), and multiple-case glioma families may disproportionately include relatives with meningioma and melanoma (Paunu et al., 2002). However, meningioma risk was not related to a family history of melanoma in a Swedish record-linkage study (Hemminki et al., 2001b). Some multiple-case melanoma families have inherited mutations in the P16INK4 or P14ARF splice variants of the CDKN2A gene (Bahuau et al., 1998; Randerson-Moor et al., 2001), which also is inactivated in a high proportion (25%–46%) of sporadic atypical/anaplastic meningiomas (Bostrom et al., 2001; Weber et al., 1997). Meningioma risk was not increased among individuals who had a first-degree relative with breast cancer in a Swedish investigation (Hemminki et al., 2001a). However, in several studies, women have had a 1.4- to 3.5-fold excess risk of a second primary breast cancer after meningioma (Helseth et al., 1989), or second primary meningioma after breast cancer (Malmer et al., 2000), or both (Custer et al., 2002; Schoenberg et al., 1975). Explanations posited for the increased risks include treatment for the first tumor or increased diagnostic surveillance or autopsy in cancer survivors (Adami et al., 1984). Neither explanation can account for our observation of an increased risk of meningioma among younger individuals (ages 18–49) who had a family history of breast cancer, compared with those of similar age who did not. The increased risk supports the possibility of shared genetic or environmental factors that contribute to risk of both tumors. However, among older individuals (aged ≥50), we observed a reduced risk of meningioma among those with a family history of breast cancer, based on a small number of cases (n = 3). While early-onset cancer may be more likely to be genetically determined, our age-specific findings could be due to chance. The association between meningioma and breast cancer, and the increased risks of both among Jewish women (Inskip et al., 2003; Preston-Martin 1989), might suggest assessment of the familial breast cancer genes BRCA1 and BRCA2. In one study, 60 sporadic meningiomas from an unselected general hospitalized population did not demonstrate loss of heterozygosity or altered mRNA expression of either gene (Kirsch et al., 1997).
The strengths and limitations of the study should be taken into account in interpretation of our findings. Our study benefited from rapid case identification through regional referral hospitals, which facilitated prompt interviews: 91% of meningioma cases and 82% of schwannoma cases were interviewed within 3 weeks of the qualifying diagnosis. However, many of the increased risks that we observed were based on a small number of observations and, considered in light of the number of hypotheses tested, could have arisen by chance. We had very limited statistical power to detect whether personal or family history of cancer was related to schwannoma risk (40% power to detect a 2-fold and 80% power to detect a 3-fold increased risk of schwannoma for a 5% prevalence of a particular family history of cancer). Brain tumor patients may have been more likely to recall brain or other tumors in relatives than controls. We did not validate the familial or personal history of cancers or benign tumors reported by the respondent, but 75% and 84% of familial cancer reports were confirmed by medical records in two glioma case-control studies (Airewele et al., 1998; Wrensch et al., 1997).
The etiology of meningioma and schwannoma is difficult to investigate, as they are relatively uncommon and historically have not been required to be reported to most cancer registries, and the asymptomatic nature of some tumors means that a proportion is detected incidentally. Thus, the low recorded incidence is probably an underestimate, and factors associated with tumor development may be confused with those related to diagnosis. Our study results require replication and suggest that investigations of families that also include other benign brain/CNS tumors, melanoma, or breast cancer might shed further light regarding shared environmental or genetic factors that contribute to meningioma etiology.
We are grateful to Emily Khoury, Brian Paul, Patsy Thompson, Donna Houpt, Kelli Williamson, Timothy Wilcosky, Sandra McGuire, Renee Karlsen, Patricia Yost, Janice Whelan, Douglas Watson, Diane Fuchs, Bob Saal, and Christel McCarty for their valuable assistance during the conduct of the study.
2Abbreviations used are as follows: CI, confidence interval; ICDO-2, International Classification of Diseases for Oncology, 2nd edition; NF2, neurofibromatosis type 2; OR, odds ratio; SIR, standardized incidence ratio.